The Expanded Groove MIDI Dataset (E-GMD) is a large dataset of human drum performances, with audio recordings annotated in MIDI. E-GMD contains 444 hours of audio from 43 drum kits and is an order of magnitude larger than similar datasets. It is also the first human-performed drum transcription dataset with annotations of velocity. It is based on our previously released Groove MIDI Dataset.

Contents

Dataset

This dataset is an expansion of the Groove MIDI Dataset (GMD). GMD is a dataset of human drum performances recorded in MIDI format on a Roland TD-11 electronic drum kit. To make the dataset applicable to ADT, we expanded it by re-recording the GMD sequences on 43 drumkits using a Roland TD-17. The kits range from electronic (e.g., 808, 909) to acoustic sounds. Recording was done at 44.1kHz and 24 bits and aligned within 2ms of the original MIDI files.

We maintained the same train, test and validation splits across sequences that GMD had. Because each kit was recorded for every sequence, we see all 43 kits in the train, test and validation splits.

Split Unique Sequences Total Sequences Duration (hours)
Train 819 35,217 341.4
Test 123 5,289 50.9
Validation 117 5,031 52.2
Total 1,059 45,537 444.5

Given the semi-manual nature of the pipeline, there were some errors in the recording process that resulted in unusable tracks. If your application requires only symbolic drum data, we recommend using the original data from the Groove MIDI Dataset.

For more information about how the dataset was created and several applications of it, please see the paper where it was introduced: Improving Perceptual Quality of Drum Transcription with the Expanded Groove MIDI Dataset.

Download

E-GMD is provided as a zip file containing the MIDI and WAV files as well as metadata in CSV format. A MIDI-only archive of the dataset is also available.

The metadata file has the following fields for every MIDI/WAV pair (same as GMD, with the addition of kit_name):

Field Description
drummer An anonymous string ID for the drummer of the performance.
session A string ID for the recording session (unique per drummer).
id A unique string ID for the performance.
style A string style for the performance formatted as “<primary>/<secondary>”. The primary style comes from the Genre List below.
bpm An integer tempo in beats per minute for the performance.
beat_type Either “beat” or “fill”
time_signature The time signature for the performance formatted as “<numerator>-<denominator>”.
midi_filename Relative path to the MIDI file.
audio_filename Relative path to the WAV file (if present).
duration The float duration in seconds (of the MIDI).
split The predefined split the performance is a part of. One of “train”, “validation”, or “test”.
kit_name Name of the drum kit on the Roland TD-17 used for synthesis.

V1.0.0

e-gmd-v1.0.0.zip

Size: 90GB (132GB uncompressed)
SHA256: 7d9a264fb4c9eabd9fec09d5f8e333192f529b1a1b845d170279a977ac436053

e-gmd-v1.0.0-midi.zip

Size: 103MB (164MB uncompressed)
SHA256: 5e70a6f4d760385a5e5ec986a2f02179d96f61181a920e592876b577a75844d3

Metadata files:

License

Creative Commons License

The dataset is made available by Google LLC under a Creative Commons Attribution 4.0 International (CC BY 4.0) License.

How to Cite

If you use the E-GMD dataset in your work, please cite the paper where it was introduced:

Lee Callender, Curtis Hawthorne, and Jesse Engel. "Improving Perceptual Quality
  of Drum Transcription with the Expanded Groove MIDI Dataset." 2020.
  arXiv:2004.00188.

You can also use the following BibTeX entry:

@misc{callender2020improving,
    title={Improving Perceptual Quality of Drum Transcription with the Expanded Groove MIDI Dataset},
    author={Lee Callender and Curtis Hawthorne and Jesse Engel},
    year={2020},
    eprint={2004.00188},
    archivePrefix={arXiv},
    primaryClass={cs.SD}
}

Please also make sure to specify which version of the dataset you are using.